In epithelial cells alternative splicing of fibroblast growth factor receptor-2 (FGFR2) transcripts leads to the expression of the FGFR2(IIIb) isoform, whereas in mesenchymal cells the same process results in the synthesis of FGFR2(IIIc), which include exon IIIc. This regulated splicing is disrupted during progression of prostate carcinomas leading to the inappropriate expression of FGFR2 isoforms. To visualize the use of FGFR2 exon IIIc in Dunning AT3 tumors in syngeneic rats we constructed minigene constructs that report on alternative splicing. Imaging these alternative splicing decisions revealed unexpected mesenchymal-epithelial transitions (MET) in these primary tumors. The understanding of this epithelial-mesenchymal plasticity, revealed by differences in splicing, is the overarching goal of this application. To this end we propose to accomplish the following specific aims: Specific Aim 1. To characterize the connections between FGFR2 splicing and epithelial-mesenchymal plasticity in prostate tumors. We have developed a new bichromatic reporter that can be used to detect both MET and epithelial-mesenchymal transitions (EMT). We have also extended our observations to well characterized human prostate tumor cells in culture. Work accomplished in this aim will test the hypothesis that epithelial plasticity is a general property of aggressive prostate carcinomas in humans. Specific Aim 2. To characterize the epithelial-mesenchymal plasticity in primary prostate tumors revealed by alterations in splicing. In this aim we will characterize the molecular signatures and behavior of the MET clusters in primary tumors. Specific aim 3. To characterize the connections between epithelial-mesenchymal plasticity and metastatic behavior. We will investigate the connections between epithelial-mesenchymal plasticity and metastatic behavior. Finally, we will isolate circulating prostate tumor cells from both the AT3 rat tumors and patients, and we will determine their FGFR2 splicing status and their epithelial state. Work accomplished in aim two and in this aim will test the hypothesis that epithelial-mesenchymal plasticity leads to increased tumor fitness. The significance of this work is twofold: First we will unravel basic information about gene expression programs controlled by alternative splicing during prostate tumor progression. Second, the potential significance to human health is underscored by the magnitude of prostate cancer as a health problem. Identifying the tumors/cells that will progress to metastatic disease is a challenge and a priority. The work proposed here should identify useful markers and perhaps even therapeutic targets.